A hybrid of back propagation neural network and genetic algorithm for optimization of injection molding process parameters

This paper presents a hybrid optimization method for optimizing the process parameters during plastic injection molding (PIM). This proposed method combines a back propagation (BP) neural network method with an intelligence global optimization algorithm, i.e. genetic algorithm (GA). A multi-objective optimization model is established to optimize the process parameters during PIM on the basis of the finite element simulation software Moldflow, Orthogonal experiment method, BP neural network as well as Genetic algorithm. Optimization goals and design variables (process parameters during PIM) are specified by the requirement of manufacture. A BP artificial neural network model is developed to obtain the mathematical relationship between the optimization goals and process parameters. Genetic algorithm is applied to optimize the process parameters that would result in optimal solution of the optimization goals. A case study of a plastic article is presented. Warpage as well as clamp force during PIM are investigated as the optimization objectives. Mold temperature, melt temperature, packing pressure, packing time and cooling time are considered to be the design variables. The case study demonstrates that the proposed optimization method can adjust the process parameters accurately and effectively to satisfy the demand of real manufacture.     

Sisal-glass fiber hybrid biocomposite: Optimization of injection molding parameters using Taguchi method for reducing shrinkage

The current work presents an application of Taguchi method to optimize injection molding (IM) process parameters of sisal-glass fiber hybrid biocomposite. Six parameters that influence flow and cross-flow shrinkage such as injection pressure, melt temperature, mold temperature, holding pressure, cooling time and holding time were selected as variables and two hybrid biocomposites were used with different content of sisal (SF) and glass fiber (GF); SF20GF10 and SF10GF20. For the experimental design, L18 orthogonal array with a mixed-level design and signal-to-noise (S/N) of smaller-the-better was used. Optimal combination IM parameters were determined and the significant variables were identified using ANOVA. Optimized flow and cross-flow shrinkage values for SF20GF10 were 0.53% and 0.85% and the values for SF10GF20 were 0.47% and 0.88% respectively. Comparison was made with the shrinkage requirements of an automotive material specification suggesting that hybrid biocomposites with optimized IM parameters meet the dimensional requirements of automotive parts.     

基于代理模型和改进遗传算法的注塑翘曲优化

提出一种最小化制品翘曲的注塑工艺参数优化集成方法.以空调柜机顶盖注塑制品开发为例,该方法使用Moldflow软件分析制品的翘曲变形,运用田口方法确定与制品翘曲量密切相关的工艺因素,然后采用响应曲面法(RSM)和改进的精英保留自适应遗传算法(EAGA)相结合的方法,建立主要影响工艺参数与制品翘曲量之间的关系模型,通过对模型寻优以实现对制品翘曲的优化.该方法的适用性在制品的实际生产中得到了验证.     

An artificial neural network approach to multiphase continua constitutive modeling

Constitutive equations describe intrinsic relationships among sets of material system parameters. This study utilizes artificial neural networks in place of a traditional micromechanical approach to calculate the global (macroscopic) elastic properties of composite materials given the local (microscopic) properties and local geometry. This approach is shown to be more computationally efficient than conventional numerical micromechanical approaches. An eight sub-celled representative volume element is used for the local geometry. Multi target artificial neural networks (MTANNs) and single target artificial neural networks are studied for applicability in predicting the global properties. The best performing MTANN achieves a precision of 9%. The single target artificial neural networks (STANNs) perform best and predicts the global properties within a target error of 5.3%. The computation time is 1.8 s for all six STANNs to predict six global properties for 19,683 different microstructures.     

Influence of fabric shear and flow direction on void formation during resin transfer molding

In resin transfer molding, void type defect is one of common process problems, it degenerates the mechanical performances of the final products seriously. Void content prediction has become a research hotspot in RTM, while the void formation when the flow direction and the tow direction are not identical or the fabric is sheared has not been studied to date. In this paper, based on the analysis of the resin flow velocities inside and outside fiber tows, a mathematical model to describe the formation of micro- and meso-scale-voids has been developed. Particular attention has been paid on the influence of flow direction and fabric shear on the impregnation of the unit cell, so their effects on the generation and size of voids have been obtained. Experimental validation has been conducted by measuring the formation and size of voids, a good agreement between the model prediction and experimental results has been found.     

Creep failure time prediction of polymers and polymer composites

A theoretical approach for the prediction of creep rupture time of polymers and polymer composites is analyzed in the present work. This analysis takes into account the viscoelastic path at small strains and the viscoplastic path at higher stresses. The calculation of the rate of creep strain is based on a thermally activated rate process, while the emergence and growth of plastic strain, with increasing creep time, is also taken into account. When the accumulated strain attains values, high enough to lead to failure, its slope versus time exhibits an abrupt change. At this specific time, the creep rate function in respect to time appears a minimum. The creep failure time is defined as the time where the creep rate takes its minimum value. The model has been tested for various types of polymeric materials, as well as for polymer composites. Once the model parameters are estimated from short time creep strain data, then it was proved to successfully predict the creep failure time at a variety of stress levels, for all material types examined.     

Numerical prediction of saturation in dual scale fibrous reinforcements during Liquid Composite Molding

This paper presents a fractional flow model based on two-phase flow, resin and air, through a porous medium to simulate numerically Liquid Composites Molding (LCM) processes. It allows predicting the formation, transport and compression of voids in the modeling of LCM. The equations are derived by combining Darcy’s law and mass conservation for each phase (resin/air). In the model, the relative permeability and capillary pressure depend on saturation. The resin is incompressible and the air slightly compressible. Introducing some simplifications, the fractional flow model consists of a saturation equation coupled with a pressure/velocity equation including the effects of air solubility and compressibility. The introduction of air compressibility in the pressure equation allows for the numerical prediction of the experimental behavior at low constant resin injection flow rate. A good agreement was obtained between the numerical prediction of saturation in a glass fiber reinforcement and the experimental observations during the filling of a test mold by Resin Transfer Molding (RTM).     

Multiscale constitutive modeling for plastic deformation of nanocrystalline materials

Macroscopic and microscopic constitutive modeling that can display large plastic deformation with shear band were presented for nanocrystalline materials subjected to uniaxial load over a wide strain rate range. The macroscopic model implemented with parameters microscopic meaning was established based on the theory of plastic dissipation energy. The microscopic model based on deformation mechanisms was composed of two parts: hardening and softening stages. In the hardening stage, the phase mixture model was used and a shear band deformation mechanism was proposed in the softening stage. Numerical simulations shown that the predications were in good agreement with experimental data. Finally, a parameter of normalized softening rate was proposed and its characteristics were evaluated quantitatively. It can be concluded that the failure strain could be prolonged when the normalized softening rate decrease through changing the softening path.     

Experimental study of warpage and shrinkage in injection molding of HDPE/rPET/wood composites with multiobjective optimization

In this paper, injection molding of squared parts with 1.25 mm in thickness, composed of wood plastic composites (high-density polyethylene, recycled polyethylene terephthalate, and wood flour), was done. The warpage and volumetric shrinkage in the parts was determined experimentally with various process conditions (packing time, melt temperature, wood content, and packing pressure). The experiments were done based on Box–Behnken design of experiments. The significance of each parameter and model was evaluated by analysis of variance (ANOVA). ANOVA showed that packing time and melt temperature are the most significant parameters on warpage and wood content is the most significant on volumetric shrinkage. Packing pressure and wood content had no considerable effect on warpage and packing time on shrinkage too. To obtain optimal process conditions for minimum warpage and shrinkage, a multiobjective optimization based on Pareto front was developed. Response surface method was used to find the relationships between input parameters and objective functions, and genetic algorithm presented the Pareto front solutions to determine the optimum solution. It was observed that there was a good agreement between the predicted optimum values and the experiments.     

Coupling filtration and flow during liquid composite molding: Experimental investigation and simulation

Molding composites constituted of fiber reinforcements, resin and fillers is of prime interest for many transportation applications. Dealing with the flow of particle-filled resin in a fibrous network raises the issue of particle retention and viscosity increase. The present study aims at simulating such molding through an efficient coupling between a filtration model, that has been previously described, and a flow model (Darcy’s law). First, filling experiments are realized so as to separate cases: cake filtration, retention and no retention for two types of single-scale porous materials (polyester felt and glass fiber mat) injected with a resin filled with micro-beads. Then results of filler content, viscosity, permeability, pressure, retention profiles are simulated from the coupling between filtration and flow. Experimental data of filler profiles in the final parts, resin flow front evolution and injection times are compared with predictions obtained from the simulation.     

Robust recurrent neural network modeling for software fault detection and correction prediction

Software fault detection and correction processes are related although different, and they should be studied together. A practical approach is to apply software reliability growth models to model fault detection, and fault correction process is assumed to be a delayed process. On the other hand, the artificial neural networks model, as a data-driven approach, tries to model these two processes together with no assumptions. Specifically, feedforward backpropagation networks have shown their advantages over analytical models in fault number predictions. In this paper, the following approach is explored. First, recurrent neural networks are applied to model these two processes together. Within this framework, a systematic networks configuration approach is developed with genetic algorithm according to the prediction performance. In order to provide robust predictions, an extra factor characterizing the dispersion of prediction repetitions is incorporated into the performance function. Comparisons with feedforward neural networks and analytical models are developed with respect to a real data set.     

Particle filtration and distribution during the liquid composite molding process for manufacturing particles containing composite materials

The effects of major process parameters on particle filtration and distribution were investigated by using newly developed microscopic methodology using an electron probe micro-analyzer and numerical simulation. The mapping results indicated that well-dispersed particles were distributed uniformly in the inter-tow and intra-tow regions. Agglomerates were likely to be filtered at the boundary or inside of the fiber bundle. The results of quantitative analyses showed particle concentrations in the inter-tow region to be uniform throughout the composite part, whereas the intra-tow concentrations varied according to particle size and fiber orientation. The poor dispersion state of the CNT-Ag particles resulted in quite irregular distributions. A high volume fraction of the fiber preform resulted in a lower particle concentration inside the fiber tow. Numerical analysis of the filtration of large clusters of particles indicated that filtration occurred in the initial stage of the injection process at the tow boundaries.     

The wear rates and performance of three mold insert materials

In this study, a rapidly solidified aluminum alloy was compared with beryllium copper and 6061 aluminum alloys in terms of their wear rates, hardness and performance as mold insert materials. A Vickers hardness measuring machine and a tribometer were used to determine the hardness values and wear rates of the materials. Three sets of mold inserts were made of these materials, and the insert surfaces and the molded plastic lens surfaces were characterized using a scanning electron microscope and a surface profilometer, respectively. The investigation results indicate that the BeCu alloy has the lowest wear rate, while aluminum 6061-T6 has the highest wear rate. Although the rapidly solidified aluminum alloy is not as hard as the BeCu alloy, the differences between their wear rates and hardness values are not as great as the differences between aluminum 6061-T6 and the BeCu alloy. The results also indicate that the rapidly solidified aluminum alloy performs much better than aluminum 6061-T6 in molding of plastic lenses and is comparable to the BeCu alloy. It is able to attain finer surfaces of the molded plastic lenses. This is an important finding, and this means that the rapidly solidified aluminum alloy can replace the BeCu alloy as a good mold insert material, because beryllium (Be) is a toxic element. The finding gives the industry a better choice for selection of mold insert materials.     

The limit drawing ratio and formability prediction of advanced high strength dual-phase steels

Dual phase (DP) steels, being among advanced high-strength steels (AHSS), have been successfully used in the sheet metal stamping of automotive components for its great benefit in reducing vehicle weight while improving car safety. In their practical application, one of the major challenges is related to formability prediction of onset crack. This paper first conducts limiting drawing ratio (LDR) experiments to identify the maximum blank diameter of SPFC340, DP600, DP800 and DP1000 with onset crack. The fracture modes of these four types of blanks are then compared and classified into two categories: necking crack and shear crack. Further for DP1000, appropriate hardening formula is determined to fit the flow curve derived by the tensile test. Three yielding models (Hill-48, Batlat-89 and Banabic-2005) are compared with each other in the numerical simulation of DP1000 LDR onset crack. The investigation shows that a Swift and Hockett–Sherby combined formula is in good agreement with the flow curve of the tensile test and Batlat-89 yield model successfully predicts the onset shear crack of DP AHSS.     

Transient gas flow technique for inspection of fiber preforms in resin transfer molding

A transient gas flow method was developed to determine the quality of fibrous preforms in resin transfer molding (RTM) prior to resin injection. The method aims at detecting defects resulting from preform misplacement in the mold, accidental inclusions, preform density variations, race tracking, shearing, etc. Unlike the previously developed method based on steady-state gas flow, the new method allows for the acquisition of continuous time-varying pressure data from multiple ports during a single test. The validity of the method was confirmed by one-dimensional flow experiments.     

Multiscale modeling of unsaturated flow in dual-scale fiber preforms of liquid composite molding III: reactive flows

The woven, stitched or braided fabrics used in liquid composite molding (LCM) display partial saturation behind moving flow-front in an LCM mold which is caused by delayed impregnation of fiber tows. In this part 3 of the present series of three papers, a novel multiscale approach proposed in parts 1 and 2 [1] and [2] is adapted for modeling the unsaturated flow observed in the dual-scale fabrics of LCM under non-isothermal, reactive conditions. The volume-averaged species or resin cure equation, in conjunction with volume-averaged mass, momentum and energy (temperature) equations, is employed to model the reactive resin flow in the inter-tow (gap) and intra-tow (tow) regions with coupling expressed through several sink and source terms in the governing equations. A coarse global-mesh is used to solve the global (gap) flow over the entire domain, and a fine local mesh in form of the unit-cell of periodic fabrics is employed to solve the local (tow) flows. The multiscale algorithm based on the hierarchical computational grids is then extended to solve the dual-scale flow under reactive conditions. The simulation is compared with a two-color experiment and a previously published two-layer model. Significant differences between the temperatures and cures of the gap and tow regions of the dual-scale porous medium are observed. The ratio of pore volumes in the tow and gap regions, the effective thermal conductivity in the tows, and the reaction rate are identified as the important parameters for temperature and cure distributions in the gap and tow regions.     

Development of slate fiber reinforced high density polyethylene composites for injection molding

During the last decade the use of fiber reinforced composite materials has consolidated as an attracting alternative to traditional materials due to an excellent balance between mechanical properties and lightweight. One drawback related to the use of inorganic fibers such as those derived from siliceous materials is the relative low compatibility with conventional organic polymer matrices. Surface treatments with coupling agents and the use of copolymers allow increasing fiber–matrix interactions which has a positive effect on overall properties of composites. In this research work we report the use of slate fiber treated with different coupling agents as reinforcement for high density polyethylene from sugarcane. A silane (propyltrimethoxy silane; PTMS) and a graft copolymer (polyethylene-graft-maleic anhydride; PE-g-MA) were used to improve fiber–matrix interactions on HDPE-slate fiber. The effect of the different compatibilizing systems and slate fiber content were evaluated by scanning electron microscopy (SEM), dynamic thermomechanical analysis (DTMA) as well as mechanical properties (tensile, flexural and impact). The results show that the use of silane coupling agents leads to higher fiber–matrix interactions which has a positive effect on overall mechanical properties. Interesting results are obtained for composites containing 30 wt.% slate fiber previously treated with propyltrimethoxy silane (PTMS) with an increase in tensile and flexural strength of about 16% and 18% respectively.     

Materials modeling and simulation of isothermal forging of rolled AZ31B magnesium alloy: Anisotropy of flow

Isothermal forging of a rib–web shape in AZ31B magnesium alloy in the rolling direction was conducted at speeds of 0.01–10 mm s⁻¹ in the temperature range of 300–500 °C with the purpose of validating the results of materials models involving kinetic analysis and processing map. The process was also simulated using finite element method DEFORM to obtain the local values of strain and strain rate. Forging parallel to the rolling direction in the range 375–550 °C and 0.0003–0.3 s⁻¹ under the conditions of dynamic recrystallization (DRX) resulted in a symmetrical cup-shape while at other conditions an elliptical boat-shape was produced with the major axis coinciding with the transverse direction and the minor axis aligned with the normal direction. This anisotropy of flow has been attributed to the strong basal texture in the rolled plate and the dominance of prismatic slip at lower temperatures. In the DRX domain on the other hand, pyramidal slip dominates along with cross-slip as the recovery mechanism, which destroys the initial texture and restores the symmetry of flow. The grain size variation for forgings done in the DRX domain validated the predictions of the material models.     

A semi-analytical simulation strategy and its application to warpage of autoclave-processed CFRP parts

The manufacturing of composite structures is accompanied by fabrication induced deformations. Those deformations are undesirable and lead to transgression of geometric tolerances in the finished parts. In order to get the part within aspired dimensional tolerances, geometrical compensation of the tool is necessary. This often iterative conducted tooling-rework is commonly time consuming and costly. This paper presents an shell element based. semi-analytical simulation approach focusing on warpage deformations due to tool part interaction, in order to account for manufacturing induced deformations within the tool design process. Deviation measurements on test specimen level serve as inputs for the calculation of equivalent coefficients of thermal expansion according to the proposed analytical model. Thus, ‘warpage properties’ of different prepreg – tool–material combinations are determined. The use and the practicability of the developed approach is demonstrated by means of the calculation of a warpage compensated tool surface.     

Synthesis and characterization of new composites: PANI/Na-AlSBA-3 and PANI/Na-AlSBA-16

The new aluminosilicate materials (Na-AlSBA-3 and Na-AlSBA-16) were synthesized for application in the preparation of composites. Silica mesoporous materials were obtained following the sol-gel method and post-synthesis alumination. These methods were effective for the synthesis of SBA-3 and SBA-16, showing XRD patterns and other characteristics in agreement with the literature.Aniline-saturated hosts were prepared by adsorption of aniline (exposed to the equilibrium vapors from liquid aniline) into the mesoporous materials. Polyaniline/Na-AlSBA-3 (PANI-3) and polyaniline/Na-AlSBA-16 (PANI-16) composites have been synthesized by an in situ polymerization of aniline-saturated hosts. TG, FTIR, XRD, SEM and TEM were used to characterize the resulting composites. These studies show that PANI is generated inside the channel of the hosts. PANI-16 has an amount of emeraldine salt higher than PANI-3 composite. The electrical conductivity measurements confirmed that PANI and mesoporous materials were true hybrid nanocomposites. The conductive properties of these composites were compared with those of other composites (polyaniline/Na-AlMCM-41 and polyaniline/Na-AlSBA-15) reported.